Flotillin-2 Deficiency Leads to Reduced Lung Metastases in a Mouse Breast

SHORT COMMUNICATION Flotillin-2 deﬁciency leads to reduced lung metastases in a mouse breast cancer model

T Berger1,9, T Ueda1,2,9, E Arpaia1, IIC Chio1, EA Shirdel1,3, I Jurisica4,3,5, K Hamada6, A You-Ten1, J Haight1, A Wakeham1, CC Cheung7 and TW Mak1,8,4

Flotillin microdomains, specialized lipid raft domains in cell membranes, serve as physical platforms for many different molecules important in crucial intracellular signaling pathways. Flotillin-2 (Flot2), together with flotillin-1, is a marker for lipid raft microdomains distinct from caveolar lipid rafts, and has been implicated in the progression of cancer and metastasis formation. Based largely on studies in xenograft models, flotillin-2 has been implicated in the progression of multiple types of human tumors, including breast cancer. In our studies, we identified flotillin-2 as highly amplified in a high-throughput comparative genomic hybridization screen of human breast cancer cell lines and breast tumor samples. Short hairpin RNA-mediated reduction of flotillin-2 protein levels significantly reduced the tumorigenicity and metastatic capability of a human breast cancer cell line in vivo. We generated mice deficient for flotillin-2 and also found a reduction of flotillin-1 protein levels and complete absence of flotillin-specific membrane microdomains in these mice. To examine the role of Flot2 in mammary tumorigenesis and lung metastasis, we used an in vivo molecular genetics approach, crossing a well-characterized transgenic mouse model of breast cancer, the MMTV-PyMT (mouse mammary tumor virus-polyoma middle T antigen) mouse, with gene-targeted Flot2 À / À mice. Flotillin-2 deficiency lead to a striking reduction in the number of lung metastasis observed, but had no influence on primary tumor formation in this model. Our results indicate, using a novel in vivo animal model approach, that Flot2 is an important regulator of mammary tumor-derived lung metastasis.

Oncogene (2013) 32, 4989–4994; doi:10.1038/onc.2012.499; published online 12 November 2012 Keywords: breast cancer; metastasis; MMTV-PyMT; ﬂotillin-1; reggie1/2

INTRODUCTION important for extracellular matrix penetration, depends on lipid 13 Breast cancer is the most frequently diagnosed cancer in women rafts, and lipid rafts are concentrated in the leading edge of 14 worldwide, and metastatic breast cancer is the leading cause of breast cancer cell invadopodia. Flotillin-1 expression correlates death from cancer among women globally.1 The most frequent with increasing grade in human breast cancer samples,15 and sites of breast cancer metastasis are bone, lung, liver and brain. knockdown of flotillin-1 impairs breast cancer cell proliferation Both breast cancer progression and metastasis formation depend and tumorigenicity.15 Enhanced Flot2 expression is associated in part on abnormalities of lipid rafts, due to their crucial role in a with melanoma progression in humans16 and with metastasis number of cellular mechanisms that are dysregulated in breast formation in human head and neck squamous cell carcinomas.17 tumor cells, such as altered protein signaling and trafficking.2 Lipid In mouse xenografts, Flot2 overexpression has been reported to rafts are plasma membrane microdomains that are highly drive the transformation of a non-tumorigenic, non-metastatic enriched in cholesterol and glycosphingolipids,3,4 and serve as melanoma cell line (SB2) into a highly tumorigenic, metastatic cell platforms for many crucial intracellular signaling proteins.2 Two line.16 In cell cultures, Flot2 knockdown reduces cell spreading, different types of lipid rafts exist: the caveolae, which are highly whereas epidermal growth factor-dependent phosphorylation of enriched in proteins from the caveolin family,5,6 and non-caveolar Flot2 by Src kinase enhances cell spreading.9 lipid rafts, which are characterized by their expression of flotillin-1 In this study, we performed a high-throughput comparative and flotillin-2, Flot2 (also known as reggie-2 and reggie-1, genomic hybridization screen of human breast cancer cell lines respectively7). Flotillins were initially thought to be associated and breast tumor samples to identify genomic regions showing with caveolae,8 but it was later shown that lipid rafts containing amplification across both. We found that the Flot2 gene encoding flotillin microdomains were distinct from caveolae.9,10 Flot2 was highly amplified in several breast cancer cell lines and The signaling pathways associated with both types of lipid rafts breast tumor samples. On the basis of these findings and the are altered in various kinds of cancer cells.11,12 In particular, the reported association of Flot2 with breast tumorigenesis and formation of invadopodia, which are membrane protrusions metastasis, we employed two genetic approaches to investigate

1The Campbell Family Institute for Breast Cancer Research, University Health Network, Toronto, ON, Canada; 2Department of Disease Model, Research Institute for Radiation Biology and Medicine, Hiroshima University, Hiroshima, Japan; 3Techna Institute, University Health Network and IBM Life Sciences Discovery Centre, Toronto ON, Canada; 4Departments of Medical Biophysics and Immunology, University of Toronto, Toronto, ON, Canada; 5Department of Computer Science, University of Toronto, Toronto ON, Canada; 6Okada Projects at the Center for AIDS Research, Kumamoto University, Kumamoto, Japan; 7Department of Pathology, University Health Network, Toronto, ON, Canada and 8Ontario Cancer Institute, University Health Network, Toronto, ON, Canada. Correspondence: Professor TW Mak, The Campbell Family Institute for Breast Cancer Research, University Health Network, 620 University Avenue, Suite 706, Toronto, ON, Canada M5G 2M9. E-mail: [email protected] 9These authors contributed equally to this work. Received 8 August 2012; revised 14 September 2012; accepted 18 September 2012; published online 12 November 2012 Flotillin-2 regulates breast cancer metastasis T Berger et al 4990

Figure 1. Flot2 is essential for breast cancer cell proliferation in vivo.(a) Copy number alterations in a panel of 50 breast cancer cell lines (a kind gift from Dennis Slamon, UCLA, CA, USA) and 17 breast cancer tissues (a kind gift from Susan Done, UHN, ON, Canada) as assessed by array comparative genomic hybridization. Construction of the sub-megabase resolution tiling (SMRT) array used for these experiments is described in de Leeuw et al.25 and Ishkanian et al.,26 and the methods are described in more detail in Watson et al.27 The BAC clone containing Flot2 was within a longer 17q11.2 amplicon (human genome version hg17) and was amplified to levels greater than log2 ratio (tumor/normal) of 0.3 in four breast cancer tumor tissues and six breast cancer cell lines. In this panel, four patient samples harboring the Flot2 amplicon are shown in yellow. The Flot2 amplicon resides upstream from the Her2 amplicon (pink) and the HoxB amplicon (green) on chromosome 17. (b) Immunoblot confirmation of the stable short hairpin RNA-mediated knockdown of Flot2 in two independent MDA-MB-231-luc cell lines. Short hairpin RNAs against Flot2 were purchased ready for lentiviral expression from Open Biosystems Products (Huntsville, AL, USA). Lentiviral particles were produced in 293FT cells and virus-containing culture supernatants were used to infect cells in the presence of 8 g/ml polybrene (Sigma-Aldrich, Oakville, ON, Canada). Levels of flotillin-1 and Flot2 protein in whole-cell lysates of empty vector control cells (ev) and shFlot2- 10 and shFlot2-12 cells10,12 were determined using primary mouse anti-flotillin-1 and anti-Flot2 monoclonal antibodies (both from BD Transduction Laboratories, Mississauga, ON, Canada). Actin (loading control) was detected using rabbit anti-actin antibody (Sigma-Aldrich, Oakville, ON, Canada). Bands were visualized using the appropriate secondary antibodies and enhanced chemiluminescence (ECL-kit, Amersham Biosciences, Piscataway, NJ, USA) according to the manufacturer’s recommendations. (c–e) Knockdown of Flot2 inhibits the tumorigenicity of breast cancer cells in vivo. The mammary fat pads of athymic nude NIH-III mice were xenografted with shFlot2-10 (n ¼ 9), shFlot2-12 (n ¼ 9), or empty vector control, (n ¼ 5) cells (1 Â 106 cells per mouse). The total tumor burden, as assessed by total tumor volume in cm3 (c) or total tumor wet weight in grams (d), was measured at the experimental end point (50 days). (e) Representative mammary tumors recovered at 50 days from mice xenografted with control MDA-MB-231-luc cells (Loc) or shFlot2-10 or shFlot2-12 cells. (f, g) Knockdown of Flot2 leads to delayed lung engraftment of mammary tumor cells and prolonged survival in severe combined immunodefecient mice. (f) Control MDA-MB-231-luc cells or shFlot2-10 or shFlot2-12 cells (5 Â 105 in 100 ml phosphate-buffered saline) were injected into the tail veins of severe combined immunodefecient mice (n ¼ 8/group). In situ imaging of engrafted cells at 19 days post injection was performed using a Xenogen apparatus (IVIS Imaging Systems, Caliper LifeSciences, Hopkinton, MA, USA). (g) Survival of the mice in (g) over 40–50 days was evaluated using Kaplan–Meier methodology. All animal experiments were approved by the Animal Care Committee of the University Health Network (Toronto, ON, Canada). For c, d, the data were analyzed using the Student’s t-test and are presented as the mean±s.d. Values of Po0.05 were considered statistically significant; **Po0.01, ***Po0.001. For g, the data were analyzed using the log-rank test; *Po0.05.

Oncogene (2013) 4989 – 4994 & 2013 Macmillan Publishers Limited Flotillin-2 regulates breast cancer metastasis T Berger et al 4991 the role of Flot2 in these processes. First, using gene silencing in a human breast cancer cell line (MDA-MB-231-luc), we have demonstrated that Flot2 knockdown reduces the tumorigenicity of human breast cancer cells in mouse xenografts. Second, using genetic ablation of Flot2 in mice, we have shown that Flot2 deﬁciency leads to a signiﬁcant reduction in lung metastases in the well-established transgenic mouse mammary tumor virus (MMTV) polyoma middle T antigen (PyMT) model of breast cancer.18 Our results point to a crucial function for Flot2 in the metastasis of breast cancer cells.

RESULTS AND DISCUSSION We performed a comparative genomic hybridization screen of 17 human breast cancer samples and 50 human breast cancer cell lines and identified Flot2 as being located in a region on human chromosome 17q11.2 that is commonly amplified in human breast cancers. Use of the MultiExperiment Viewer version 4.8 (MeV)19,20 to examine the Flot2 amplicon in breast cancer tissues in the context of other genomic aberrations on chromosome 17 showed that although the Flot2 amplicon is upstream of both the Her2 and HoxB amplicons, it is an independent genomic aberration, with large regions of non-amplified DNA separating Figure 2. Immunoblot analysis of flotillin-1 and Flot2 protein expression. (a, b) Protein samples from (a) MEFs generated from the duplication events (Figure 1a). þ / þ À / À To investigate the role of Flot2 in the tumorigenic activity of Flot2 and Flot2 mice, and (b) the indicated organs from these animals, were analyzed by immunoblotting using anti-flotillin-1 breast cancer cells, we generated two stable Flot2 knockdown or anti-Flot2 antibodies. Actin, loading control. Results are repre- MDA-MB-231-luc cell lines, designated shFlot2-10 and shFlot2-12 sentative of three mice per group. (c, d). Flot2-deficient mice lack (Figure 1b). The proliferation of these cells in vitro was not flotillin microdomains. Detergent-resistant membranes were isolated inhibited by their lack of Flot2 (Supplementary Figure S1), in from extracts of Flot2 þ / þ and Flot2 À / À MEFs by fractionation using contrast to the reported poor proliferation of flotillin-1 knockdown detergent resistant membranes (DRM) floatation sucrose density- breast cancer cells.15 To determine whether Flot2 depletion had gradient centrifugation as previously described.21 Fractions including an effect on the tumorigenicity of breast cancer cells in vivo,we membranes (fractions 8–10, boxed) were analyzed by immunoblotting xenografted wild-type (WT) MDA-MB-231-luc cells, or shFlot2-10 or to detect (c) flotillin-1 and (d) caveolin-1 using specific primary shFlot2-12 MDA-MB-231-luc cells, into the mammary fat pads of antibodies from Santa Cruz Biotechnology, (Santa Cruz, CA, USA). Briefly, cell pellets were resuspended in 1 ml ice-cold TNE buffer athymic NIH-III nude mice. As shown in Figures 1c–e, total tumor (50 mM Tris pH 7.4, 150 mM NaCl, 5 mM EDTA) containing a protease volumes and weights were significantly reduced in mice bearing inhibitor cocktail (Roche Canada, Laval, QC, Canada), and 1 ml ice- Flot2-depleted MDA-MB-231-luc cells compared with mice bear- cold 2% Triton X-100 in TNE was added for a 30 min incubation on ing WT MDA-MB-231-luc cells. These results indicate that Flot2 ice. After pelleting of nuclei by centrifugation at 1000 g for 5 min, indeed is required for breast tumorigenesis in vivo. 2 ml 80% sucrose in TNE was added to adjust the lysate to 40% We next tested the ability of Flot2 knockdown cells to survive in sucrose. The lysate was pipetted into the bottom of an ultracentrifuge the circulation and extravasate into the stroma of the lung and tube (Beckman Instruments, Palo Alto, CA, USA), and 6 ml 35% sucrose form metastases. WT MDA-MB-231-luc cells, or shFlot2-10 or and then 2 ml 5% sucrose (both in TNE) were carefully layered on top. shFlot2-12 MDA-MB-231-luc cells, were injected intravenously into The gradient was centrifuged at 30 000 r.p.m. (SW40 rotor, Beckman Coulter, Mississauga, ON, Canada) for 20.6 h. Twelve 1 ml fractions the tail veins of severe combined immunodefecient mice. Mouse were collected (starting from the bottom) by tube puncture and survival was monitored and engraftment of tumor cells into the proteins were precipitated using MeOH/CHCl3. lungs was evaluated by xenogen imaging. After 19 days, mice engrafted with WT MDA-MB-231-luc cells showed a strong accumulation of tumor cells in the lungs, whereas mice engrafted (Supplementary Figure S2b). Four correctly targeted ES cell clones with shFlot2-10 MDA-MB-231-luc cells showed barely detectable were injected (separately) into C57BL/6J blastocysts to generate lung engraftment (Figure 1f). As a consequence, mice injected chimeric mice. F1 heterozygote mice originating from one clone with shFlot2-10 MDA-MB-231-luc or shFlot2-12 MDA-MB-231-luc were intercrossed to produce the Flot2 þ / þ and Flot2 À / À F2 cells had significantly increased lifespans (about 44 days) progeny analyzed in this report. Immunoblotting of extracts of compared twith mice that had been engrafted with WT MDA- mouse embryonic fibroblasts (MEFs) prepared from Flot2 þ / þ and MB-231-luc cells (about 36 days; Figure 1g, log-rank test, Po0.05). Flot2 À / À littermate mice confirmed that the knockout Flot2 allele Thus, mice engrafted with breast cancer cells in which Flot2 was a null mutation, and that no Flot2 protein was detectable in expression is substantially depleted show delayed lung metastasis these cells (Figure 2a). formation. Interestingly, the expression of flotillin-1 protein was also To be able to utilize mouse mammary tumorigenesis models reduced in Flot2 À / À MEFs (Figure 2a), a finding confirmed in that better reflect the complex multistage nature of human several organs of these mutants (Figure 2b). These results are in tumorigenesis, we used homologous recombination in mouse accordance with a previous report showing that genetic ablation embryonic stem (ES) cells to disrupt the Flot2 gene and generate of flotillin-1 concomitantly reduces Flot2 expression,21 but are in Flot2 À / À knockout mice completely lacking Flot2 protein contrast to our results obtained with short hairpin RNA-mediated expression (Supplementary Figure S2a). The targeting vector was Flot2 protein reduction, where we did not observe a significant designed to replace most of the Flot2 coding region, including flotillin-1 protein level decrease. It is feasible that although both exons 3–10, with the PGK-neo cassette. We obtained 13 flotillin-1 and -2 are necessary for the formation of flotillin independent ES cell clones carrying the mutated Flot2 allele, membrane microdomains, the residual amount of Flot2 still and Southern blot analysis of their DNA hybridized to a 30 flanking present in the Flot2 knockdown MDA-MB-231-luc cell lines is probe or a neomycin-specific probe confirmed a single integration sufficient to stabilize flotillin-1 and prevent its degradation.

& 2013 Macmillan Publishers Limited Oncogene (2013) 4989 – 4994 Flotillin-2 regulates breast cancer metastasis T Berger et al 4992

Figure 3. Flot2 deficiency significantly reduces the formation of lung metastases in the MMTV-PyMT breast cancer mouse model. (a–d) Primary tumor formation in MMTV-PyMT mice is not affected by deletion of Flot2. (a) Age at time of killing due to allowable tumor burden was recorded for each Flot2 þ / þ PyMT (n ¼ 26) and Flot2 À / À PyMT (n ¼ 28) mouse. (b–d) The number of mammary glands with grossly visible tumors (b), the total volume in cm3 of all mammary tumors (c) and the total wet weight in grams of all mammary tumors (d) were measured for each mouse in a.(e, f) Lung metastasis is significantly reduced in Flot2 À / À PyMT mice. (e) The number of macroscopically visible lung metastases was counted for each mouse in a.(f) The number of microscopically visible metastases was counted in five hematoxylin–eosin- stained sections of lung (taken 200 mm apart) for each mouse in a.Fora–f, data were analyzed using the Student’s t-test and are presented as the mean±s.d. Values of Po0.05 were considered statistically significant; ns, not significant; *Po0.05; **Po0.01.

Lipid rafts are insoluble in cold non-ionic detergents,8 which generated Flot2 þ / þ MMTV-PyMT (Flot2 þ / þ PyMT) and Flot2 À / À / offers a means of isolating them. We prepared cold Triton X-100- MMTV-PyMT (Flot2 À À PyMT) mice and compared spontaneous solubilized extracts of Flot2 þ / þ and Flot2 À / À MEFs and tumor formation and metastasis in female animals. Both subjected them to sucrose density-gradient ultracentrifugation. Flot2 À / À PyMT (n ¼ 28) and Flot2 þ / þ PyMT (n ¼ 26) mice were Consistent with our previous finding, no residual flotillin-1 protein killed at about 160 days after birth (Figure 3a), a time point shortly was detected in the detergent-resistant fractions of the mutant before the tumor burdens in both strains of mice reached the extracts (Figure 2c). In contrast, the distribution of caveolin was humane end point. We analyzed numbers of mammary glands not altered in fractionated extracts of Flot2 À / À MEFs (Figure 2d). bearing tumors (as assessed by gross histological analysis), the These results indicate that a complete ablation of Flot2 abolishes total volume of all mammary tumors per animal, and the total wet the formation of functional flotillin, but not caveolar membrane weight of all mammary tumors per animal, but no significant microdomains. differences were detected (Figures 3b–d). These data indicate that To analyze the role of Flot2 in tumor initiation, progression and a lack of Flot2 does not interfere with overall primary tumor metastasis, we took advantage of the MMTV-PyMT mouse model of formation in MMTV-PyMT mice. mammary tumorigenesis. WT female MMTV-PyMT mice develop A great advantage of the MMTV-PyMT mouse model is that it multifocal dysplastic atypia in the mammary epithelium as early as can be seamlessly used to study both primary mammary tumor 3 weeks after birth. These lesions are thought to be the origin of development and metastasis. A high percentage of MMTV-PyMT the primary mammary adenocarcinomas that appear in these mice develop metastases in the lungs, and these lesions can be animals with an average latency of about 35–40 days.18 Thereafter, readily observed macroscopically and microscopically.18 To these primary tumors spontaneously metastasize to the lungs with determine whether Flot2 influences metastasis formation, we a high penetrance. The relevance of the MMTV-PyMT model to examined the lungs of Flot2 þ / þ PyMT and Flot2 À / À PyMT mice at human breast cancer has been validated by both genetic and the time of killing (160 days). Upon gross observation to detect histological studies,22,23 and the stages of tumorigenesis in MMTV- surface lesions, we found that the lungs of Flot2 À / À PyMT mice PyMT mice demonstrate a striking similarity to the stages of human exhibited 10-fold fewer visible macro-metastases compared with ductal adenocarcinogenesis. In particular, both ErbB2/Neu and the lungs of Flot2 þ / þ PyMT controls (2.1±1.1 vs 21.7±7.2; cyclin D1 expression are upregulated in MMTV-PyMT tumor cells, Figure 3e). We then prepared a minimum of five lung tissue just as that occurs in human ductal adenocarcinoma cells.22,24 sections (200 mm apart) per lung, stained them with hematoxylin– To test whether Flot2 is an oncogene contributing to the eosin, and counted numbers of visible micro-metastases. formation of mammary tumors in the MMTV-PyMT model, we Again, there was a significant decrease in the number of

Oncogene (2013) 4989 – 4994 & 2013 Macmillan Publishers Limited Flotillin-2 regulates breast cancer metastasis T Berger et al 4993 Importantly, however, the number of macro-metastases visible on the lungs of NIH-III mice engrafted with Flot2 À / À PyMT cells was significantly reduced compared with mice engrafted with Flot2 þ / þ PyMT cells (0.9±0.7 vs 6.4±1.4; Figure 4c). Further- more, the number of lung micro-metastases detectable by hematoxylin–eosin staining was five-fold lower in NIH-III mice that had been engrafted with Flot2 À / À PyMT cells compared with those engrafted with Flot2 þ / þ PyMT cells (4.2±2.6 vs 20.1±4.5; Figure 4d). The results of our experiments using genetic ablation of Flot2 and a well-established mouse model of mammary tumorigenesis and metastasis show for the first time that Flot2 has an important role in regulating breast cancer-derived lung metastasis formation. These data confirm previously reported findings associating Flot2 with metastasis formation in human head and neck squamous cell carcinomas17 and the transformation of a non-metastatic melanoma cell line into a metastatic cell line,16 and underline the importance of future research into the function of flotillin microdomains in cancer progression and metastasis formation.

CONFLICT OF INTEREST The authors declare no conﬂict of interest.

ACKNOWLEDGEMENTS We thank Lily Zhou for expert technical assistance, and Dr Mary Saunders for insightful scientiﬁc editing. This work was supported by grant no. 179815 from the Canadian Institutes of Health Research to TWM. This research was funded in part by the Ontario Ministry of Health and Long Term Care. The views expressed do not necessarily reﬂect those of the OMOHLTC. ES is in part supported by Ontario Research Fund (GL2-01-030) to IJ, and IJ was in part supported by Canada Research Chair Program.

Figure 4. Flot2 deficiency leads to significantly fewer lung metas- REFERENCES tases in an MMTV-PyMT xenograft model. (a) Immunoblot con- 1 Boyle P, Levin B. International Agency for Research on Cancer. World Health firmation of Flot2 protein deficiency in a primary mammary tumor Organization. World Cancer Report 2008. International Agency for Research on / / cell line (Flot2 À À PyMT cells) generated in-house from Flot2 À À Cancer; Distributed by WHO Press: Lyon Geneva, 2008, 510pp. 28 PyMT mice using a previously described protocol. The immunoblot 2 Simons K, Toomre D. Lipid rafts and signal transduction. Nat Rev Mol Cell Biol / analysis of the corresponding Flot2 þ þ PyMT control cell line is 2000; 1: 31–39. also shown. (b–d) Decreased lung metastasis after xenografting. 3 Le Moyec L, Tatoud R, Eugene M, Gauville C, Primot I, Charlemagne D et al. Cell (b) The mammary fat pads of athymic nude NIH-III mice were injected and membrane lipid analysis by proton magnetic resonance spectroscopy in five / / with Flot2 þ þ PyMT cells (n ¼ 11) or Flot2 À À PyMT cells (n ¼ 13) breast cancer cell lines. Br J Cancer 1992; 66: 623–628. 6 (1 Â 10 cells per mouse). The total tumor burden, as assessed by 4 Nohara K, Wang F, Spiegel S. Glycosphingolipid composition of MDA-MB-231 and tumor weight in grams, was measured in each mouse at the MCF-7 human breast cancer cell lines. Breast Cancer Res Treat 1998; 48: 149–157. experimental end point (50 days post injection). No differences in 5 Rothberg KG, Heuser JE, Donzell WC, Ying YS, Glenney JR, Anderson RG. Caveolin, primary tumor formation were detected. (c) The number of a protein component of caveolae membrane coats. Cell 1992; 68: 673–682. macroscopically visible lung metastases was counted for each 6 Yamada E. The fine structure of the renal glomerulus of the mouse. J Biophys mouse in b.(d) The number of microscopically visible metastases Biochem Cytol 1955; 1: 551–566. was counted in five hematoxylin–eosin-stained sections/lung (taken 7 Schulte T, Paschke KA, Laessing U, Lottspeich F, Stuermer CA. Reggie-1 and 200 mm apart) for each mouse in b.Forb–d, data were analyzed using reggie-2, two cell surface proteins expressed by retinal ganglion cells during axon the Student’s t-test and are presented as the mean±s.d. Values of regeneration. Development 1997; 124: 577–587. Po0.05 were considered statistically significant; **Po0.01. 8 Bickel PE, Scherer PE, Schnitzer JE, Oh P, Lisanti MP, Lodish HF. Flotillin and epidermal surface antigen define a new family of caveolae-associated integral membrane proteins. J Biol Chem 1997; 272: 13793–13802. micro-metastases in the lungs of Flot2 À / À PyMT mice compared 9 Neumann-Giesen C, Fernow I, Amaddii M, Tikkanen R. Role of EGF-induced þ / þ ± ± tyrosine phosphorylation of reggie-1/flotillin-2 in cell spreading and signaling to with Flot2 PyMT controls (10.8 4.0 vs 82.3 31.8; Figure 3f). the actin cytoskeleton. J Cell Sci 2007; 120(Pt 3): 395–406. These results clearly point to the involvement of Flot2 in the 10 Stuermer CA, Langhorst MF, Wiechers MF, Legler DF, Von Hanwehr SH, Guse AH formation of lung metastases in the MMTV/PyMT model. et al. PrPc capping in T cells promotes its association with the lipid raft To further analyze the role of Flot2 in metastasis, we derived proteins reggie-1 and reggie-2 and leads to signal transduction. FASEB J 2004; 18: individual cell lines from mammary tumors of Flot2 À / À PyMT and 1731–1733. Flot2 þ / þ PyMT mice. Immunoblotting confirmed the lack of Flot2 11 de Laurentiis A, Donovan L, Arcaro A. Lipid rafts and caveolae in signaling by protein in the cell line derived from Flot2 À / À PyMT tumor growth factor receptors. Open Biochem J 2007; 1: 12–32. (Figure 4a). These cells (or their control counterparts) were then 12 Pike LJ. Lipid rafts: bringing order to chaos. J Lipid Res 2003; 44: 655–667. xenografted into the mammary fat pads of NIH-III nude 13 Buccione R, Caldieri G, Ayala I. Invadopodia: specialized tumor cell structures for the focal degradation of the extracellular matrix. Cancer Metastasis Rev 2009; 28: mice (n ¼ 11/13 per group). By 50 days post injection, the total À / À 137–149. wet weight of tumors arising from the Flot2 PyMT cell line 14 Yamaguchi H, Takeo Y, Yoshida S, Kouchi Z, Nakamura Y, Fukami K. Lipid rafts and was equivalent to the total wet weight of tumors arising from caveolin-1 are required for invadopodia formation and extracellular matrix / the Flot2 þ þ PyMT cell line (3.5±0.4 vs 2.5±0.3 g; Figure 4b). degradation by human breast cancer cells. Cancer Res 2009; 69: 8594–8602.

& 2013 Macmillan Publishers Limited Oncogene (2013) 4989 – 4994 Flotillin-2 regulates breast cancer metastasis T Berger et al 4994 15 Lin C, Wu Z, Lin X, Yu C, Shi T, Zeng Y et al. Knockdown of FLOT1 impairs cell 22 Lin EY, Jones JG, Li P, Zhu L, Whitney KD, Muller WJ et al. Progression to malig- proliferation and tumorigenicity in breast cancer through upregulation of nancy in the polyoma middle T oncoprotein mouse breast cancer model provides FOXO3a. Clin Cancer Res 2011; 17: 3089–3099. a reliable model for human diseases. Am J Pathol 2003; 163: 2113–2126. 16 Hazarika P, McCarty MF, Prieto VG, George S, Babu D, Koul D et al. Up-regulation 23 Qiu TH, Chandramouli GV, Hunter KW, Alkharouf NW, Green JE, Liu ET. Global of Flotillin-2 is associated with melanoma progression and modulates expression expression profiling identifies signatures of tumor virulence in MMTV-PyMT- of the thrombin receptor protease activated receptor 1. Cancer Res 2004; 64: transgenic mice: correlation to human disease. Cancer Res 2004; 64: 5973–5981. 7361–7369. 24 Maglione JE, Moghanaki D, Young LJ, Manner CK, Ellies LG, Joseph SO et al. 17 Rickman DS, Millon R, De Reynies A, Thomas E, Wasylyk C, Muller D et al. Transgenic polyoma middle-T mice model premalignant mammary disease. Prediction of future metastasis and molecular characterization of head and neck Cancer Res 2001; 61: 8298–8305. squamous-cell carcinoma based on transcriptome and genome analysis by 25 de Leeuw RJ, Davies JJ, Rosenwald A, Bebb G, Gascoyne RD, Dyer MJ et al. microarrays. Oncogene 2008; 27: 6607–6622. Comprehensive whole genome array CGH profiling of mantle cell lymphoma 18 Guy CT, Cardiff RD, Muller WJ. Induction of mammary tumors by expression of model genomes. Hum Mol Genet 2004; 13: 1827–1837. polyomavirus middle T oncogene: a transgenic mouse model for metastatic 26 Ishkanian AS, Malloff CA, Watson SK, DeLeeuw RJ, Chi B, Coe BP et al. A tiling disease. Mol Cell Biol 1992; 12: 954–961. resolution DNA microarray with complete coverage of the human genome. Nat 19 Saeed AI, Bhagabati NK, Braisted JC, Liang W, Sharov V, Howe EA et al. TM4 Genet 2004; 36: 299–303. microarray software suite. Methods Enzymol 2006; 411: 134–193. 27 Watson SK, deLeeuw RJ, Ishkanian AS, Malloff CA, Lam WL. Methods for high 20 Saeed AI, Sharov V, White J, Li J, Liang W, Bhagabati N et al. TM4: a free, open- throughput validation of amplified fragment pools of BAC DNA for constructing source system for microarray data management and analysis. BioTechniques 2003; high resolution CGH arrays. BMC Genomics 2004; 5:6. 34: 374–378. 28 Arpaia E, Blaser H, Quintela-Fandino M, Duncan G, Leong HS, Ablack A et al. The 21 Ludwig A, Otto GP, Riento K, Hams E, Fallon PG, Nichols BJ. Flotillin microdomains interaction between caveolin-1 and Rho-GTPases promotes metastasis by con- interact with the cortical cytoskeleton to control uropod formation and neutrophil trolling the expression of alpha5-integrin and the activation of Src, Ras and Erk. recruitment. J Cell Biol 2010; 191: 771–781. Oncogene 2012; 31: 884–896.

Supplementary Information accompanies the paper on the Oncogene website (http://www.nature.com/onc)

Oncogene (2013) 4989 – 4994 & 2013 Macmillan Publishers Limited